Anthony Tether said he has reinvigorated DARPA's innovative spirit in his nearly eight years as head of the agency. Critics say they hope President-elect Barack Obama appoints a new DARPA chief, claiming Tether has cut back basic research, a charge he refutes.

In the past eight years, "there was tremendous deterioration in the agendas of specific agencies and DARPA is a prime example," said Edward Lazowska, a past chairman of the Computing Research Association and a computer science professor at the University of Washington. "DARPA's focus became transitioning existing research results into products for the military," he added.

"That's not true at all, it's an urban legend," said Tether.

DARPA's annual spending on basic research has more than doubled from less than $90 million in 2001 to more than $200 million today. During that period DARPA's overall budget has only grown by 50 percent from $2 billion to $3 billion, Tether said.

The heads of computer science departments at top universities including MIT and Stanford shared their own numbers in a report published by EE Times in November 2005.

Bill Dally, chairman of Stanford's computer science department, said in that report the amount of the university's computer research funded by DARPA dropped from 80 percent in the late 1990's to just 15 percent. At MIT, DARPA funding made up 62 percent of its computer science budget with the money aimed at 3-5 year projects in 1999, but it has declined to about 24 percent and aimed at shorter term projects, said Rodney Brooks, former chairman of MIT's computer science department, speaking in the 2005 article.

Tether said he did not dispute those numbers, but they were not due to declines in DARPA funding for basic research so much as a lack from those universities of fresh ideas the agency was interested in funding.

"A lot of places that are getting mature like computer science departments feel like they have an entitlement, but we don't have any entitlements, we just fund good ideas," said Tether.

"While computer science was the first field to go public with its concerns about the changing nature of DARPA, other fields are feeling the same effects," said Lazowska. "There is nothing specific to computer science about the changes that DARPA has undergone under its current leadership," he added.

In an informal discussion, two Berkeley researchers noted DARPA has not spent some of annual allocated funds for research. Tether said the agency has indeed returned as much as $150 million a year to the Department of Defense for other uses when budgeted research projects failed to meet their milestones.

"It would be amazing if all our projects passed all their milestones," said Tether. "I would love to have anyone--especially the people at Berkeley--tell me they were working on such great ideas that they always work," he added.